(C) PLOS One [1]. This unaltered content originally appeared in journals.plosone.org.
Licensed under Creative Commons Attribution (CC BY) license.
url:https://journals.plos.org/plosone/s/licenses-and-copyright

------------



Inducible CRISPR activation screen for interferon-stimulated genes identifies OAS1 as a SARS-CoV-2 restriction factor

['Oded Danziger', 'Department Of Microbiology', 'Icahn School Of Medicine At Mount Sinai', 'New York', 'United States Of America', 'Roosheel S. Patel', 'Emma J. Degrace', 'Mikaela R. Rosen', 'Brad R. Rosenberg']

Date: 2022-06

Interferons establish an antiviral state through the induction of hundreds of interferon-stimulated genes (ISGs). The mechanisms and viral specificities for most ISGs remain incompletely understood. To enable high-throughput interrogation of ISG antiviral functions in pooled genetic screens while mitigating potentially confounding effects of endogenous interferon and antiproliferative/proapoptotic ISG activities, we adapted a CRISPR-activation (CRISPRa) system for inducible ISG expression in isogenic cell lines with and without the capacity to respond to interferons. We used this platform to screen for ISGs that restrict SARS-CoV-2. Results included ISGs previously described to restrict SARS-CoV-2 and novel candidate antiviral factors. We validated a subset of these by complementary CRISPRa and cDNA expression experiments. OAS1, a top-ranked hit across multiple screens, exhibited strong antiviral effects against SARS-CoV-2, which required OAS1 catalytic activity. These studies demonstrate a high-throughput approach to assess antiviral functions within the ISG repertoire, exemplified by identification of multiple SARS-CoV-2 restriction factors.

To counteract viral infection, interferon induces the expression of antiviral effectors collectively termed Interferon Stimulated Genes (ISGs). Different effector genes restrict different viruses through a variety of mechanisms. Identifying the particular ISGs that restrict specific viruses can uncover viral vulnerabilities and inform therapeutic strategies. Here, we developed a CRISPR-activation (CRISPRa) experimental strategy for screening hundreds of pooled ISGs for antiviral activity in parallel while accounting for the ability of some ISGs to inhibit cell cycle and to promote cell death. We applied this approach to detect ISGs that restrict SARS-CoV-2, the virus that causes COVID-19. In a series of controlled experiments, we identified multiple ISGs that counteract SARS-CoV-2 infection, and further validated a subset of these ISGs through focused CRISPRa and cDNA expression studies. Our results validate several previously identified ISGs, and identify multiple novel antiviral effectors with activity against SARS-CoV-2. Overall, our work demonstrates an experimental system for the controlled assessment of ISG antiviral activities, and expands the present understanding of innate antiviral defense against SARS-CoV-2.

Funding: This work was supported in part by National Institutes of Health grants R01 AI151029 (B.R.R.) and U01 AI150748 (B.R.R.), and funds from the Icahn School of Medicine at Mount Sinai (B.R.R.). Research reported in this paper was also supported by the Office of Research Infrastructure of the National Institutes of Health under award numbers S10OD018522 and S10OD026880. The content is solely the responsibility of the authors and does not necessarily represent the official views of the funders, including the National Institutes of Health. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Here, we report an ISG-focused CRISPRa screen to identify ISGs that modulate SARS-CoV-2 infection in lung epithelial cells. To mitigate the potential antiproliferative and/or proapoptotic effects of certain ISGs that could impact their library representation, we engineered a Doxycycline (Dox)-inducible CRISPRa system that enables precise temporal control of ISG induction. Using a pooled screen strategy, we tested more than 400 ISGs for effects on SARS-CoV-2 infection in both wildtype cells and isogenic cells engineered to be insensitive to IFN. High ranking antiviral ISG hits included SARS-CoV-2 restriction factors previously identified in recent screens (LY6E [ 26 ], CD74 [ 27 ], TRIM25 [ 30 ] and ERLIN1 [ 29 ]). We identified and validated previously undescribed antiviral roles for additional ISGs such as CTSS (Cathepsin S). In agreement with Wickenhagen et al. and Soveg et al., [ 31 , 32 ] we also identified OAS1 (2′-5′-oligoadenylate synthetase 1) as a SARS-CoV-2 restriction factor capable of inhibiting viral infection and the generation of progeny virus. Taken together, our findings demonstrate the utility of a novel inducible CRISPRa platform for antiviral genetic screens and identify multiple ISGs capable of restricting SARS-CoV-2.

CRISPR-activation (CRISPRa), in which guide RNA (gRNA)-directed endonuclease-deficient Cas9 along with transcriptional activators are targeted to a gene of interest (GOI) to induce expression, offers an alternative to cDNA ectopic expression screens. Advantages include easy to produce gRNA libraries, physiologically relevant expression levels, and multiplexing capabilities [ 42 ]. Gene transcription is initiated from endogenous promoters, and therefore has the potential to generate multiple isoforms for genes with alternative splicing programs. Studies in which genome-wide libraries of activating gRNAs uncovered important host factors in viral infection systems provide an important proof of concept to the characterization of ISGs using CRISPRa [ 43 – 45 ]. Importantly, two recent preprints report genome-wide CRISPRa screens for genes with antiviral potential against SARS-CoV-2 [ 46 , 47 ].

Studies evaluating the antiviral potential of individual ISGs are often implemented through the ectopic expression of ISG cDNA libraries followed by viral challenge to screen for those ISGs that confer resistance [ 3 , 4 , 29 , 33 , 34 ]. While effective in identifying the antiviral potential of many ISGs, arrayed ISG cDNA screens are not without drawbacks including limited throughput, technically demanding cloning and validation of individual expression constructs, and high costs. Ectopic cDNA overexpression can be prone to artifactual expression patterns or functions [ 35 , 36 ]. In addition, cDNA expression screen libraries typically include only one isoform per gene, and therefore may overlook isoform-specific antiviral activities, as have been described for many ISGs [ 27 , 37 – 41 ].

The specific mechanisms by which IFN restricts SARS-CoV-2 have not been fully characterized. To date, of the hundreds of ISGs, only a handful have been found to restrict SARS-CoV-2 infection in different in vitro systems. Lymphocyte antigen 6 complex, locus-E (LY6E) was identified as a SARS-CoV-2 restriction factor in ISG ectopic cDNA expression screens [ 26 ]. A transposon-mediated screen identified the MHC-II invariant chain CD74 as a block to SARS-CoV-2 entry [ 27 ]. In addition, overexpression of BST2 (encoding Tetherin), an anti-HIV effector that prevents the release of nascent virions [ 28 ], has also been found to restrict SARS-CoV-2 by impeding virion release [ 29 ]. TRIM25, an interferon-induced E3 ubiquitin ligase that enhances antiviral responses downstream of RIG-I, has been shown to interact specifically with SARS-CoV-2 RNA and thereby reduce infection [ 30 ]. During preparation of this manuscript, two additional studies employing different expression systems and screening methodologies highlighted the prominent role of OAS1 in restricting SARS-CoV-2 [ 31 , 32 ].

A robust IFN response is critical for host defense against novel respiratory viruses to which immune memory from prior exposure has not been established [ 12 ]. As such, multiple lines of evidence have implicated IFN as a key component of the host response to Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), the etiologic agent of COVID-19. Although IFN can effectively block SARS-CoV-2 infection in vitro [ 13 – 16 ], the activity of IFN in different physiological contexts is more complex. Characterizations of the IFN response to SARS-CoV-2 infection in a variety of model systems suggest that SARS-CoV-2 infection may not elicit robust IFN production and ISG expression, but can induce high levels of proinflammatory cytokines [ 17 ]. These findings are consistent with single cell RNA sequencing (scRNA-seq) immune profiling studies that have reported inflammatory gene signatures and less robust IFN/ISG expression in immune cells from individuals infected with COVID-19 as compared to individuals infected with Influenza [ 18 ]. However, transcriptomic analyses of bronchoalveolar lavage fluid (BALF) from COVID-19 patients detected a strong ISG signature along with proinflammatory cytokine gene expression in immune cells [ 19 ]. Dysregulation of the IFN response in SARS-CoV-2 infection can be driven by both direct viral antagonism of innate immune mechanisms, as well as by host characteristics such as age, genetics and other comorbidities [ 20 ]. Additionally, several studies have identified autoantibodies against IFN as a significant negative survival factor for severe COVID-19, further emphasizing the prominent role of IFN in SARS-CoV-2 pathogenesis and clearance [ 21 – 24 ]. Importantly, IFN-mediated effects may also be detrimental to COVID-19 outcomes. For example, disruption of the lung epithelium by type III IFN-dependent processes has been hypothesized to expose patients to secondary infections by opportunistic bacteria [ 25 ]. Taken together, while IFN and the downstream expression of ISGs can functionally restrict SARS-CoV-2, the site(s), cell types, amount, and timing of IFN production and response play critical roles in COVID-19 pathogenesis and outcomes.

Interferons (IFN) act as key mediators of the host response to viral pathogens by establishing a general antiviral state through the coordinated induction of hundreds of interferon stimulated genes (ISGs) in infected and uninfected “bystander” cells [ 1 ]. ISGs encode functionally diverse gene products, including antiviral effectors that antagonize distinct steps of viral life cycles [ 2 ], but the antiviral mechanisms of most individual ISGs remain unknown. Studies that have systematically characterized the effects of single ISGs have demonstrated that a limited number of individual ISGs, primarily transcription factors and DNA/RNA sensors, can broadly restrict infection by multiple viruses upon overexpression in target cells [ 3 – 7 ]. Other individual ISGs have been found to restrict or even to enhance the replication of specific viruses. In addition to direct antiviral effectors, the ISG repertoire also includes genes that induce antiproliferative and/or proapoptotic programs in response to viral infection or DNA damage, thereby limiting viral spread and impeding oncogenesis [ 8 – 11 ].

Results

Inducible CRISPRa system in A549 cells We developed an optimized platform for pooled, positive selection ISG screens by adapting the well-established SunTag CRISPRa technology [48], and engineering it into A549 lung adenocarcinoma cell lines, a widely employed model for respiratory virus infection [49–51]. First, we reasoned that the antiproliferative and/or proapoptotic properties of some ISGs [52] could affect their relative representation in pooled libraries prior to infection experiments and/or independent of potential effects on virus susceptibility. To mitigate these effects, we reengineered the transcriptional transactivator component construct of the SunTag CRISPRa system to allow for Doxycycline (Dox)-inducible expression, and thereby Dox-inducible regulation of gRNA-targeted gene expression (Fig 1A). Next, we expected that IFN secretion in response to infection, with corresponding autocrine and paracrine induction of broad ISG expression throughout cultures, could interfere with assessments of individual CRISPRa-induced ISG effects within different cells in pooled screen experiments. Therefore, we transduced the modified SunTag components into A549ΔSTAT1 cell lines [53] defective in their capacity to respond to IFN (A549ΔSTAT1-SunTag), as well as the STAT1-competent A549 cell line (A549-SunTag, intact IFN response) from which they were derived. After selecting and expanding dual antibiotic-resistant single cell clones, we evaluated their capacity for Dox-inducible, gRNA-targeted gene expression. We transduced A549-SunTag and A549ΔSTAT1-SunTag cell lines with lentiviral constructs expressing gRNAs targeting the promoter region of MX1, a well-characterized ISG that restricts multiple viruses [54,55], as well as with non-targeting gRNA (NTG) controls. Following puromycin selection of gRNA-transduced cells, cultures were treated with Dox for 48 hours, and MX1 mRNA expression was assessed by quantitative RT-PCR (qRT-PCR). In both A549-SunTag and A549ΔSTAT1-SunTag genotypes, Dox induced robust increases in MX1 mRNA in cells expressing MX1 gRNAs (Fig 1B), while Dox treatment of cells expressing NTG or non-transduced cells exhibited minimal changes in MX1 gene expression. Comparison to cells pretreated with IFNα2b indicated that CRISPRa induced MX1 mRNA expression to similar levels (Fig 1B). Interestingly, baseline levels of MX1 mRNA exhibited higher Ct values in A549ΔSTAT1 cells compared to their STAT1-competent A549 counterparts, resulting in greater fold change values for MX1 expression upon Dox treatment (Fig 1B and 1C). This suggests that, even in the absence of exogenous IFN stimulation, some ISGs exhibit some level of constitutive STAT1-dependent transcription [56]. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 1. Inducible CRISPRa system demonstrates Dox-regulated, gRNA-specific gene expression and functional block of VSV-GFP infection in isogenic A549 and A549ΔSTAT1 cell lines. (A) Schematic of the Dox-inducible CRISPRa system for assessing antiviral ISG activities. (B) qRT-PCR analysis for MX1 mRNA in A549-SunTag and A549ΔSTAT1-SunTag cells transduced with MX1 gRNA, non-targeting gRNA (NTG), or in cells with no gRNA (NG), treated with Dox or IFN. Fold change gene expression in Doxon cells relative to Doxoff cells calculated using the ΔΔCt method with normalization to GAPDH. (C) Mean qRT-PCR threshold cycle (Ct) values for MX1 and GAPDH mRNA in A549-SunTag and A549ΔSTAT1-SunTag cells transduced with MX1 gRNA or NTG gRNA (NTG). Error bars indicate ± SD. (D) Representative flow cytometry histograms for GFP fluorescence in A549-SunTag and A549ΔSTAT1-SunTag transduced with MX1 gRNA (blue) or NTG (gray) gRNA and infected with VSV-GFP (M.O.I. = 1, 24hr). (E) Percent VSV-GFP-positive cells by flow cytometry quantified for n = 3 biological replicates. Bars represent mean ± SD of GFP positive cells across all replicates. Paired ratio Student’s t-test, *** p < 0.0005. https://doi.org/10.1371/journal.ppat.1010464.g001 Next, we evaluated the functional antiviral capacity of an ISG in our CRISPRa cells. A549-SunTag and A549ΔSTAT1-SunTag cells, transduced and selected to express MX1 gRNA, were treated with Dox and infected with a GFP-encoding Indiana vesiculovirus (VSV-GFP [57]). Flow cytometry analysis, using GFP expression as a marker for productive infection, demonstrated a near-complete block of infection in DoxOn A549-SunTag cells with active MX1 expression; DoxOff cells and cells expressing an NTG were not protected (Fig 1D and 1E). While we also observed a protective effect in A549ΔSTAT1-SunTag, a considerable fraction of cells remained susceptible to infection. These results indicate that Dox-inducible, CRISPRa-mediated ISG expression can effectively restrict viral infection. Furthermore, the restriction of VSV-GFP infection by MX1 is enhanced by additional, STAT1-dependent factors likely elicited by IFN production, further highlighting the utility of a STAT1-deficient screening platform for assessing the antiviral activity of individual ISGs. In sum, we established a functional, Dox-regulated CRISPRa system in isogenic cell lines with either intact or deficient IFN responses that effectively restricts viral infection upon induced expression of antiviral ISGs.

Inducible CRISPRa ISG screen for SARS-CoV-2 restriction factors To identify ISGs that restrict SARS-CoV-2, we conducted pooled gene activation screens in our engineered CRISPRa A549-SunTag lines. Our general screening strategy was to evaluate the potential of hundreds of individual ISGs to confer resistance to the cytopathic effects (CPE) of SARS-CoV-2. We began by introducing ACE2 expression into A549-SunTag and A549ΔSTAT1-SunTag cells to enable productive SARS-CoV-2 infection of our CRISPRa cells. Next, we conducted pilot experiments evaluating SARS-CoV-2 CPE for optimization of screen conditions to balance the strength of selective pressure with its duration for robust detection of hits [58]. A549-SunTag ACE2 and A549ΔSTAT1-SunTag ACE2 cells were transduced with expression constructs for gRNAs targeting LY6E, a known SARS-CoV-2 restriction factor [26], or NTGs. Following Dox treatment, cultures were infected with SARS-CoV-2 at a range of multiplicities of infection (M.O.I.), plates were fixed every 24 hours, and cell viability was estimated by Methylene blue assay (S1A Fig). DoxOn cultures expressing LY6E gRNAs exhibited increased viability when infected with SARS-CoV-2. Moreover, A549-SunTag ACE2 cultures exhibited less CPE than A549ΔSTAT1-SunTag ACE2 cultures, implicating additional STAT1-dependent antiviral factors. Based on these data, we approximated optimal infection conditions for ISG screens (M.O.I. = 3, harvest at 72 hours post-infection). To construct an ISG gRNA library for screening, we merged lists of ISGs tested in previous studies [3,4] with a list of genes upregulated by IFNβ treatment in A549 cells [53] (log 2 fold-change >2, adjusted p value < 0.05). To focus on antiviral effectors, we excluded known transcription factors [59], several central Pattern Recognition Receptors (PRRs), and Human leukocyte antigen (HLA) genes (S1 Table contains the full list of ISGs in the library). For each ISG, we selected 3 gRNA sequences from the optimized Calabrese CRISPRa collection [60]. ISG gRNA sequences were supplemented with an additional 24 NTG controls, for a final list of 1,266 gRNAs targeting 414 ISGs (S2 Table). Gene activation screens were conducted in both A549-SunTag ACE2 and A549ΔSTAT1-SunTag ACE2 cells, in multiple independently transduced clones, across two independent experiments. A549-SunTag ACE2 and A549ΔSTAT1-SunTag ACE2 cells were transduced with our ISG gRNA library (M.O.I. = 0.1), puromycin selected, and expanded. 48 hours prior to SARS-CoV-2 infection, appropriate cultures were treated with Dox to induce ISG expression (Fig 2A). All experiments included DoxOff and DoxOn conditions, each of which included Mock or SARS-CoV-2 infection. At 72 hours post-infection, gRNA libraries were prepared from surviving cells and sequenced to assess gRNA relative enrichment/depletion. As we observed somewhat more CPE than expected in the first experiment, we slightly relaxed the selection pressure in the second experiment (additional wash for excess virus, details in Materials and Methods). All samples passed quality control metrics for multiple parameters as previously defined [61], exhibiting a high number of reads/sample and a high ratio of mapping between reads to the gRNA library (mean = 73%, SD = ± 5%) (S1B Fig). Mapped read counts were normalized within each genotype to allow for comparative analysis of gRNA enrichment/depletion (S1C Fig). To take advantage of our Dox-inducible system and replicated design for rigorous detection of ISG effects on SARS-CoV-2, we used MAGeCK-MLE [61,62] to test differential gRNA enrichment with a linear model including factors for Dox treatment (Off/On), SARS-CoV-2 infection status (Mock/Infected), Clone (independent gRNA library transduction 1/2), and Experiment (1/2, Full screen results: S3 Table). Importantly, this analysis strategy enabled assessment of ISG effects on SARS-CoV-2 in the context of potential antiproliferative/proapoptotic ISG effects that might independently alter gRNA abundance (assessed by DoxOn Mock vs DoxOff Mock and corresponding interaction term in the model). PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 2. Inducible CRISPRa ISG screen for SARS-CoV-2 restriction factors. (A) Schematic of inducible CRISPRa ISG screens for SARS-CoV-2: Single cell clones of A549-SunTag ACE2 and A549ΔSTAT1-SunTag ACE2 cells were transduced with a library of 1,266 gRNAs and infected with SARS-CoV-2 (M.O.I. = 3, 72 hours). After onset of SARS-CoV-2 CPE, genomic DNA was extracted from surviving cells, and enriched/depleted gRNAs were evaluated by high throughput sequencing followed by MAGecK MLE analysis. (B) MAGeCK-MLE analysis of SARS-CoV-2 ISG screens in A549-SunTag ACE2 and A549ΔSTAT1-SunTag ACE2 cells. Plot depicts MAGeCK MLE β scores for Dox status coefficient (DoxOff mock infected vs DoxOn mock infected, x-axis), and SARS-CoV-2 infection status coefficient (DoxOn SARS-CoV-2 infected vs DoxOn mock infected, y-axis). ISGs passing significance selection filters (SARS-CoV-2 infection status adjusted p value and SARS-CoV-2:Dox interaction adjusted p value < 0.1) are highlighted in red/blue for “antiviral”/“proviral” effects, respectively. SARS-CoV-2 restriction factors and antiproliferative genes are labeled with gene symbols. (C) MAGeCK-MLE analysis of SARS-CoV-2 ISG screens in A549-SunTag ACE2 and A549ΔSTAT1-SunTag ACE2 cells, aligned by ISG. MAGeCK MLE β scores for SARS-CoV-2 infection status coefficient (DoxOn SARS-CoV-2 infected vs DoxOn mock infected) are plotted for each ISG (x-axis, alphabetical order); dots are sized according to significance of SARS-CoV-2:Dox interaction coefficient (-Log 10 adjusted p value) and highlighted in red/blue as in (B). Dashed line indicates ± 1 standard deviation of SARS-CoV-2 infection status coefficient β scores. (D) Comparison of candidate “antiviral”/“proviral” ISG hits passing significance filters in at least one (A549-SunTag ACE2 or A549ΔSTAT1-SunTag ACE2) genotype. Dots are shaded by ISG Z-scores for SARS-CoV-2 infection status coefficient (DoxOn SARS-CoV-2 infected vs DoxOn mock infected), and sized according to the significance of SARS-CoV-2:Dox interaction coefficient (-Log 10 adjusted p value). Filled boxes indicate ISG passing significance filters for the indicated screen. https://doi.org/10.1371/journal.ppat.1010464.g002 To identify ISGs with an effect on SARS-CoV-2 CPE in A549-SunTag ACE2 or A549ΔSTAT1-SunTag ACE2 cells, we applied a stringent two-tiered set of filters for gRNAs differentially altered by SARS-CoV-2 infection. “Antiviral” and “proviral” (as determined by MAGeCK-MLE β-score [61] for DoxOn infected vs. DoxOff infected coefficients) ISG hits were identified based on statistical significance for infection status (FDR < 0.1), while accounting for the effect of Dox treatment on gRNA abundance (linear model interaction FDR < 0.1). This strategy enabled robust identification of ISGs for which effects on SARS-CoV-2 infection were significantly stronger than their effects in mock infected cultures (e.g. due to antiproliferative, pro-apoptotic, or other effects). We detected 6 and 24 ISGs as “antiviral” (i.e. enriched by SARS-CoV-2 infection, Fig 2B and 2C) in A549-SunTag ACE2 and A549ΔSTAT1-SunTag ACE2, respectively. Conversely, we detected 10 and 2 ISGs as “proviral” (i.e. depleted by SARS-CoV-2 infection, Fig 2B and 2C) in A549-SunTag ACE2 and A549ΔSTAT1-SunTag ACE2, respectively. Antiviral hits in at least one screen (i.e A549-SunTag ACE2 or A549ΔSTAT1-SunTag ACE2) included LY6E, CD74, TRIM25 and IFITM1 (Fig 2B and 2C), each of which has been previously reported to restrict SARS-CoV-2 [26,27,30,63]. Proviral hits in at least one STAT1 genotype included CTSL (encoding Cathepsin L), an entry factor for coronaviruses [64]. Additional proviral hits CDKN1A (p21) and TNFRSF10A, ISGs with antiproliferative and/or proapoptotic effects [65,66], were depleted upon Dox treatment independently of viral infection, and were further depleted by SARS-CoV-2 infection. This is in line with previous observations that coronaviruses require cell cycle inhibition for optimal replication [67]. Interestingly, when comparing ISG hits between A549-SunTag ACE2 and A549ΔSTAT1-SunTag ACE2 screens (Fig 2D), we found only a single common antiviral hit across genotypes: OAS1 (encoding 2′-5′-oligoadenylate synthetase 1). Not surprisingly, our results suggest that STAT1-dependent transcription, perhaps in response to endogenous IFN production in A549-SunTag ACE2 screens, modulates detection of CRISPRa-induced ISG effects on SARS-CoV-2. As illustrated in Fig 2D, many of our ISG hits were similarly selected in both STAT1 genotypes (i.e. both antiviral or both proviral), but failed to clear significance thresholds in one of the screens. These discrepancies may be due to differences in relative enrichment (i.e. against all individual ISGs in the screen) of particular ISGs in IFN-responsive versus STAT1-deficient contexts. In sum, in conducting focused CRISPRa ISG screens for cell viability effects in A549 cells infected with SARS-CoV-2, we detected several previously known SARS-CoV-2 restriction factors as well as identified new candidate ISGs with putative anti-SARS-CoV-2 activities.

Comparison to other ISG-focused screens highlights common and unique antiviral hits To date, two published studies report focused ISG screens for host factors with antiviral activity against SARS-CoV-2 [29,31]. Wickenhagen et al. and Martin-Sancho et al. both employ arrayed expression strategies coupled with flow cytometry or microscopy infection readouts, respectively. Together with our present study, which utilizes a pooled expression approach, ISG-focused screens have evaluated 637 distinct genes in the context of SARS-CoV-2 infection and identified multiple antiviral hits. Each study evaluated different sets of ISGs, due to different inclusion criteria in library construction and/or the availability of sequence-verified cDNA clones for arrayed experiments. Intersecting the ISG libraries of each study revealed a set of 224 ISGs assayed in all three screens, while highlighting that each study tested distinct collections of ISGs (S2A Fig). Somewhat surprisingly, in comparing ISGs designated “antiviral” by Wickenhagen et al., Martin-Sancho et al. and our screens, we found only a single ISG (LY6E) common to all three studies (S2B Fig). A gene-level comparison of infection Z-scores, ranked from lowest (most inhibitory) to highest (least inhibitory) in each dataset, illustrated many similar trends in antiviral activities, but further exhibited the minimal overlap of significant hits across studies (S2C Fig). For example, IFITM3 was identified as a SARS-CoV-2 restriction factor by Martin-Sancho et al. [29], but did not clear significance thresholds in our screens or those reported by Wickenhagen et al. Other ISGs with evidence of restrictive activity against SARS-CoV-2, such as Z3CHAV1 (encoding ZAP) [68,69], were not reported as antiviral hits in any of the ISG screens. Such discrepancies may be attributable to differences in experimental systems (e.g. cell lines, ISG expression levels, and infection parameters). Additionally, differences from results presented here could be a result of screening methodology; unlike arrayed experiments, which evaluate the activity of individual ISGs, our pooled screening strategy evaluates relative enrichment/depletion of gRNAs, thereby highlighting the most inhibitory and enhancing ISGs in the tested set.

Validation of screen hits in targeted CRISPRa studies To confirm hits identified in the pooled screens, we first conducted “single gene” validation experiments with the CRISPRa system for a subset of candidate antiviral ISGs. A549-SunTag ACE2 and A549ΔSTAT1-SunTag ACE2 cells were transduced with expression constructs for gRNAs targeting one of eight antiviral ISG hits (CD74, LY6E, OAS1, CTSS, TRIM25, ERLIN1, ADPRHL2, GBP1) or with one of two NTGs, treated with Dox to induce gene expression, and infected with SARS-CoV-2. Compared to screening experiments (M.O.I. = 3), which were designed to select cells resistant to CPE in the context of extensive cell death, infection conditions in validation experiments were reduced slightly (M.O.I. = 2) to retain sufficient viable cells for downstream readouts. Dox-inducible, gRNA-specific CRISPRa-mediated expression for select genes (CD74, CTSS and OAS1) was confirmed in complementary qRT-PCR experiments (S3A and S3B Fig). We assessed the fraction of infected cells in DoxOn cultures (relative to fraction of infected cells in corresponding paired DoxOff cultures, set to 100%) at 24 and 72 hours post-infection. At 24 hours, the fraction of SARS-CoV-2 infected cells (evaluated by flow cytometry for SARS-CoV-2 N protein) was significantly reduced in both A549-SunTag ACE2 and A549ΔSTAT1-SunTag ACE2 cultures transduced with gRNAs targeting SARS-CoV-2 restriction factors CD74 [27] and LY6E [26] (Fig 3A and 3B). SARS-CoV-2 infection was also significantly reduced by activation of OAS1 expression in both A549-SunTag ACE2 and A549ΔSTAT1-SunTag ACE2 cultures, while TRIM25 expression demonstrated significant restriction only in A549-SunTag ACE2 cells (Fig 3A and 3B). Of note, TRIM25 enhances RIG-I signaling [70] and has been shown to interact with SARS-CoV-2 RNA [30], which may suggest that a consequent antiviral effect may require intact IFN signaling. Although some other hits exhibited consistent modest evidence of viral restriction (e.g. CTSS), no additional genes met statistical significance thresholds at the 24 hour time point. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 3. Validation of ISG screen hits in targeted CRISPRa and ectopic cDNA expression studies. (A) Representative flow cytometry histograms for SARS-CoV-2 N protein in A549-SunTag ACE2 and A549ΔSTAT1-SunTag ACE2 transduced with indicated gRNAs, treated (red) or not treated (gray) with Dox, at 24 hours post-infection with SARS-CoV-2 (M.O.I. = 2). (B) Percent of infected (SARS-CoV-2 N protein positive) cells quantified across biological replicates (n = 3, n = 2 for CD74 gRNA in A549-SunTag ACE2 cells) for experiments described in (A). Values denote percent of infected cells in Doxon cultures relative to paired Doxoff cultures. Points represent individual biological replicates, black lines indicate mean values of biological replicates for each indicated ISG gRNA. Red points indicate statistical significance (p < 0.05) as determined by paired ratio Student’s t-test. (C) Representative immunofluorescence images for SARS-CoV-2 N protein and DAPI in A549-SunTag ACE2 and A549ΔSTAT1-SunTag ACE2 transduced with indicated gRNAs and treated with Dox, at 72 hours post-infection with SARS-CoV-2 (M.O.I. = 2). (D) Percent of infected (SARS-CoV-2 N protein positive) cells quantified across biological replicates (n = 3, n = 2 for ADPRHL2 gRNA in A549-SunTag ACE cells) for experiments described in (C). Values denote percent of infected cells in Doxon cultures relative to paired Doxoff cultures. Points represent individual biological replicates, black lines indicate mean values of biological replicates for each indicated ISG gRNA. Statistical significance as in (B). (E) Representative flow cytometry histograms for SARS-CoV-2 N protein in A549 ACE2 and A549ΔSTAT1 ACE2 transduced with expression constructs for indicated cDNAs (fLUC, firefly luciferase negative control), treated (red) or not treated (gray) with Dox, at 24 hours post-infection with SARS-CoV-2 (M.O.I. = 2). (F) Percent of infected (SARS-CoV-2 N protein positive) cells quantified across biological replicates (n ≥ 3) for experiments in (E). Values denote percent of infected cells in Doxon cultures relative to paired Doxoff cultures. Points represent individual biological replicates, black lines indicate mean values of biological replicates for each indicated ISG cDNA. Red points indicate statistical significance (p < 0.05) as determined by paired ratio Student’s t-test. https://doi.org/10.1371/journal.ppat.1010464.g003 Several additional hits were confirmed to be antiviral at 72 hours post infection. Once again, the fraction of SARS-CoV-2 infected cells (here assessed by high-throughput microscopy for SARS-CoV-2 N protein due to CPE/fragile cells) was significantly reduced in A549-SunTag ACE2 cultures with activated expression of CD74, LY6E, and OAS1 (Fig 3C and 3D). We also observed significant reduction of infection in cultures expressing CTSS in both A549-SunTag ACE2 and A549ΔSTAT1-SunTag ACE2 cultures, confirming its activity as a novel SARS-CoV-2 restriction factor. At this time point, ERLIN1 expression also exhibited modest, yet significant, restriction of SARS-CoV-2 only in A549-SunTag ACE2 cells (Fig 3C and 3D). Ectopic expression of ERLIN1, a regulator of endoplasmic-reticulum-associated protein degradation (ERAD), was recently shown to restrict SARS-CoV-2 infection [29]. ADPRHL2 and GBP1, additional ISGs identified as antiviral screen hits, did not reach statistical significance in CRISPRa validation experiments (S3C and S3D Fig). Taken together, these results confirm the antiviral effects of multiple ISG screen hits in our CRISPRa system, including OAS1 and CTSS, against SARS-CoV-2.

Validation of screen hits by ectopic cDNA expression To further validate antiviral hits confirmed in CRISPRa experiments with a complementary experimental system, we tested ectopically expressed ISG cDNAs for their ability to restrict SARS-CoV-2. A549 ACE2 and A549ΔSTAT1 ACE2 cells (both without CRISPRa components) were transduced with Dox-inducible cDNA expression constructs for one of CD74, LY6E, CTSS, TRIM25, ADPRHL2 or fLuc (Firefly luciferase, negative control; additional independent experiments for OAS1 cDNAs are described in the following section). Comparison of protein expression patterns to corresponding CRISPRa cultures and assessments of Dox inducibility properties were carried out by immunoblot analysis of select screen hits (CD74 and CTSS, S4A and S4B Fig). We observed Dox-inducible, robust protein expression of target genes from both cDNA and CRISPRa; no expression was detected in cultures transduced with firefly luciferase (fLUC) cDNA or NTG. For validation experiments, after antibiotic selection and expansion, cDNA expression was induced by Dox for 48 hours, after which cells were infected with SARS-CoV-2 (M.O.I. = 2). At 24 hours post-infection, the fraction of infected cells in DoxOn and DoxOff conditions was assessed by flow cytometry for SARS-CoV-2 N protein. Dox-inducible cDNA expression of CD74, LY6E, CTSS and TRIM25 recapitulated similar patterns of SARS-CoV-2 restriction (Fig 3E and 3F) observed in the CRISPRa system (Fig 3A–3D). In addition, ADPRHL2, ADP-ribose glycohydrolase, which has not been previously described as an antiviral effector, also significantly restricted SARS-CoV-2 in cDNA expression experiments (Fig 3E and 3F). Interestingly, the apparent antiviral effect conferred by cDNA expression of CD74 was weaker than that observed in corresponding CRISPRa gRNA-directed expression experiments (Fig 3E, compared to Fig 3A). Of note, the anti-SARS-CoV-2 effects of CD74 have been ascribed to the p41 isoform [27], while the “canonical” isoform p45 (Uniprot identifier P04233-1) was tested in our cDNA experiments. Interestingly, in immunoblot analysis (S4A Fig), CRISPRa induction of CD74 from its endogenous promoter in SunTag cells resulted in predominant expression of the potent antiviral p41 isoform. However, CD74 expression from cDNA resulted in only a faint band at the expected molecular size of 45kDa but a much stronger band at 37kDa, consistent with post-translational proteolytic cleavage known to regulate the localization and activity of CD74 [71]. The soluble form of CD74 (apparent as an immunoreactive band at 25kDa [72]), was also observed in both cDNA and CRISPRa expression systems. These discrepancies likely explain the modest restriction of SARS-CoV-2 by CD74 in cDNA expression experiments as compared to CRISPRa experiments, and further highlight the potential of CRISPRa in revealing potential isoform-specific ISG antiviral effects that could be overlooked in cDNA expression studies.

Characterization of OAS1 as a potent SARS-CoV-2 restriction factor OAS1, identified as a SARS-CoV-2 restriction factor in several recent reports [31,32], was detected as an antiviral ISG hit in both A549-SunTag and A549ΔSTAT1-SunTag screens (Fig 2) and validated in targeted CRISPRa experiments at both time points tested (Fig 3). The oligoadenylate synthetase (OAS) family of ISGs are key enzymes involved in antiviral defense [1]. Upon sensing cytosolic dsRNA, OAS proteins are activated to catalyze the formation of 2′-5′-linked oligoadenylate (2′-5′A), which in turn activates latent ribonuclease L (RNaseL). Active RNaseL directly combats diverse viruses by degrading viral genomes, and indirectly by degrading cellular RNA and tRNA [73]. In humans, three members of the OAS family (OAS1, OAS2, OAS3) are capable of synthesizing 2′-5′A, and differ by size and sensitivity to dsRNA. An additional family member, OASL, is deficient in 2′-5′A catalysis but can sense dsRNA and thereby enhance RIG-I signaling [74,75]. While we detected OAS1 as a highly ranked hit in both A549-SunTag ACE2 and A549ΔSTAT1-SunTag ACE2 screens, other OAS family members were not found to confer antiviral or proviral effects (Fig 4A). PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 4. OAS1 is a potent SARS-CoV-2 restriction factor. (A) OAS family genes in inducible CRISPRa ISG screen results. For A549-SunTag ACE2 and A549ΔSTAT1-SunTag ACE2 screens, ISGs were ranked by SARS-CoV-2:Dox interaction adjusted p value (x axis). β scores for SARS-CoV-2 infection status coefficient (DoxOn SARS-CoV-2 infected vs DoxOn mock infected) are plotted on the y axis, with points passing significance filters highlighted in red/blue for “antiviral”/“proviral” effects, respectively. OAS family members are labeled as indicated. (B) Images of A549-SunTag ACE2 and A549ΔSTAT1-SunTag ACE2 cultures transduced with OAS1 gRNA or NTG, treated/not treated with Dox and infected as indicated with SARS-CoV-2 (M.O.I. = 2, 72 hours infection). Cultures were fixed and stained with Crystal violet to visualize cell viability. (C) SARS-CoV-2 growth curves measured by plaque assay. A549ΔSTAT1-SunTag ACE2 cells expressing OAS1 gRNA or NTG gRNA were infected with SARS-CoV-2 (M.O.I. = 0.1) and culture media was collected at indicated time points (x-axis, hours post infection, hpi). Titer was determined by plaque assay on Vero-E6 cells. Statistical significance at each timepoint determined by two-sided Student’s t-test, *** p < 0.0005, ** p < 0.005, * p < 0.05. (D) Immunoblot analysis of OAS1 isoforms, OAS1p46 catalytic inactive mutant, OAS2, OAS3 and OASL. Indicated genes were expressed from lentivirus vectors and gene expression was induced by Dox treatment for 48 hours prior to immunoblotting with indicated antibodies. fLUC-expressing vector was used as a negative control. (E) Representative flow cytometry histograms for SARS-CoV-2 N protein in A549 ACE2 and A549ΔSTAT1 ACE2 transduced with expression constructs for indicated OAS family member cDNAs, treated (red) or not treated (gray) with Dox, at 24 hours post-infection with SARS-CoV-2 (M.O.I. = 2). (F) Percent of infected (SARS-CoV-2 N protein positive) cells quantified across biological replicates for experiments described in (E). Values denote percent of infected cells in Doxon cultures relative to paired Doxoff cultures. Points represent individual biological replicates, black lines indicate mean values of biological replicates for each indicated cDNA. Red points indicate statistical significance (p < 0.05) as determined by paired ratio Student’s t-test. https://doi.org/10.1371/journal.ppat.1010464.g004 Given its apparent potency in both IFN-responsive and non-responsive contexts, we focused additional studies on characterizing the antiviral activities of OAS1 on SARS-CoV-2 infection. Extending the single gene CRISPRa experimental design described above (Fig 3A–3D), we assessed the effects of gRNA-activated OAS1 expression on cell viability during SARS-CoV-2 infection. Consistent with results from our positive selection cell survival screens, Dox induction of gRNA-mediated OAS1 expression conferred a dramatic improvement in cell survival assessed at 72 hours post-infection in both A549-SunTag ACE2 and A549ΔSTAT1-SunTag ACE2 cultures; DoxOff cultures and cultures transduced with NTG were readily eliminated by infection (Fig 4B). In addition to protective effects on cell viability, we also tested the direct impact of OAS1 expression on the propagation of SARS-CoV-2. To minimize the effects of endogenous IFN production on viral replication, we infected DoxOff and DoxOn A549ΔSTAT1 SunTag ACE2 cells expressing OAS1 or NTG gRNAs with a low inoculum of SARS-CoV-2 (M.O.I = 0.05), and quantified viral titers by plaque assay over time. Activation of OAS1 expression resulted in a significant decrease in the production of viral progeny, indicating that OAS1 activity functionally restricts SARS-CoV-2 infection (Fig 4C). OAS1 isoforms have been shown to differ in their antiviral activities [38]. To evaluate potential isoform-specific effects of OAS1 on SARS-CoV-2 restriction, we transduced A549 ACE2 and A549ΔSTAT1 ACE2 cells with Dox-inducible expression constructs for cDNAs encoding OAS1 canonical isoform p46 (OAS1p46, Uniprot identifier P00973-1) and a shorter OAS1 isoform (OAS1p42, Uniprot identifier P00973-2) that was recently suggested to be the highest expressed isoform of OAS1 in A549 cells [38]. To test the requirement of OAS1 catalytic activity for restricting SARS-CoV-2, we also transduced an expression construct for a catalytically inactive OAS1p46 in which amino acids 331, 332 and 333 are replaced with Alanine (OAS1p46C331A/F332A/K333A) [76]. We also included constructs for OAS2, OAS3 and OASL cDNAs. Following antibiotic selection and expansion, cDNA expression was induced with Dox and validated by immunoblotting (Fig 4D). To assay for antiviral activity, cultures were infected with SARS-CoV-2 (M.O.I. = 2) and infection was assessed by flow cytometry (SARS-CoV-2 N protein) at 24 hours post infection (Fig 4E and 4F). As expected, Dox induction of OAS1p46 expression significantly reduced the fraction of infected cells in both A549 ACE2 and A549ΔSTAT1 ACE2 cells. The p42 isoform of OAS1 exhibited similar restriction of infection. Inactivation of OAS1 catalytic activity completely ablated its antiviral effects, suggesting that the mechanism of OAS1-mediated SARS-CoV-2 restriction acts through downstream RNaseL activation. Interestingly, while expression of OAS3 exhibited some restriction of SARS-CoV-2 (significant only in A549 ACE2 cells), other OAS family members did not restrict SARS-CoV-2 infection (OAS2), or may have promoted SARS-CoV-2 infection (OASL, only in A549ΔSTAT1 ACE2 cells).

[END]

[1] Url: https://journals.plos.org/plospathogens/article?id=10.1371/journal.ppat.1010464

(C) Plos One. "Accelerating the publication of peer-reviewed science."
Licensed under Creative Commons Attribution (CC BY 4.0)
URL: https://creativecommons.org/licenses/by/4.0/


via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/

You are viewing proxied material from magical.fish. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.